Structure and function in rhodopsin: Massspectrometric identification of the abnormalintradiscal disulfide bond in misfoldedretinitis pigmentosa mutantsJohn Hwa*, Judith Klein-Seetharaman, and H. Gobind Khorana†

Departments of Biology and Chemistry, Massachusetts Institute of Technology, Cambridge, MA 02139

Contributed by H. Gobind Khorana, December 29, 2000

Retinitis pigmentosa (RP) point mutations in both the intradiscal(ID) and transmembrane domains of rhodopsin cause partial orcomplete misfolding of rhodopsin, resulting in loss of 11-cis-retinalbinding. Previous work has shown that misfolding is caused by theformation of a disulfide bond in the ID domain different from thenative Cys-110–Cys-187 disulfide bond in native rhodopsin. Herewe report on direct identification of the abnormal disulfide bondin misfolded RP mutants in the transmembrane domain by massspectrometric analysis. This disulfide bond is between Cys-185 andCys-187, the same as previously identified in misfolded RP muta-tions in the ID domain. The strategy described here should begenerally applicable to identification of disulfide bonds in otherintegral membrane proteins.

G protein-coupled receptors u signal transduction u transmembranedomain u N-ethylmaleimide u biotin-avidin affinity chromatography

A large number of naturally occurring point mutations inrhodopsin are now known. The majority of such mutations are

associated with retinitis pigmentosa (RP), and whereas they arefound in all of the three domains—cytoplasmic, transmembrane(TM), and intradiscal (ID)—those in the latter two domainsaccount for the bulk of them (1–5). Mutations in the ID domainwere found to cause partial or complete misfolding in the IDdomain, misfolding being defined by the loss of ability to bind11-cis-retinal (6–10). Studies of naturally occurring RP mutationsand of designed mutations in the ID domain showed that misfoldingwas caused by the formation of a disulfide bond different from thatin native rhodopsin. More recent work has conclusively shown thatthe abnormal disulfide bond in misfolded RP mutants in the IDdomain is between Cys-185 and Cys-187. On the one hand, theevidence came from studies of a rhodopsin mutant (in whichCys-110 was replaced by alanine) and of the RP mutants, C110F‡

and C110Y (11); on the other hand, it came from identification ofthe Cys-185–Cys-187 disulfide bond formed after reconstitution ofrhodopsin from two complementary fragments (12). Further, stud-ies of RP mutations in the TM domain of rhodopsin showed thatthey also cause misfolding by formation of an abnormal ID disulfidebond (13, 14). Identification of the abnormal disulfide bond in thelatter case acquires special significance because, if this disulfidebond is the same as that shown above for misfolded RP mutationsin the ID domain, between Cys-185 and Cys-187, then this resultwould show that defects in the packing of the helices in the TMdomain are able to cause misfolding in the ID domain. Therefore,packing of the helices in the TM domain and folding to a tertiarystructure in the ID domain must be coupled. Here we report thedirect identification by mass spectrometry of the abnormal disulfidebond in misfolded RP mutants in the TM domain (15). From theRP mutants in the TM domain studied (14), we selected four(G89D, L125R, A164V, and H211P; Fig. 1) in the present study.Misfolding in all of the four mutants is now shown to be caused bya Cys-185–Cys-187 disulfide bond. We also confirm by the method

now developed that the disulfide bond in wild-type (WT) rhodop-sin, indirectly identified previously (16–18), is indeed betweenCys-110 and Cys-187.

The aim of the strategy was to isolate and sequence peptidesadjoining the two cysteines participating in the disulfide linkage.Because of the hydrophobic nature of the opsin, the techniqueaimed at selective isolation of peptide segments adjoining thecysteines involved in the disulfide bonds (color coded in Fig. 1).In step one, all of the free cysteines in the protein werederivatized with N-ethylmaleimide (NEM). This step requireddenaturing conditions (0.5% SDS). Further, an acidic pH (pH 6)was used to avoid any disulfide-exchange reaction (19). Becauseof the low rate of reaction at pH 6, a high concentration of NEMwas used for this derivatization. The derivatized protein waspurified by binding to anti-rhodopsin 1D4-Sepharose. In steptwo, the protein, while bound to the Sepharose beads, wastreated with DTT to reduce the disulfide bond. The sulfhydrylgroups now formed were derivatized by treatment of the protein-Sepharose suspension with maleimido-butyryl-biocytin (MBB;Fig. 2). In WT rhodopsin and analogously, folded retinal-bindingrhodopsin mutants (Fig. 3A), the free sulfhydryl groups formedfrom reduction of the Cys-110 and Cys-187 disulfide bond wouldreact with the maleimido group in MBB (Cys-185 would havebeen derivatized with NEM in step one, along with all of theother sulfhydryl groups). In the misfolded proteins, the abnor-mal disulfide bond could be either between Cys-110 and Cys-185,or between Cys-185 and Cys-187. Positions for the attachment ofMBB expected for a Cys-110–Cys-185 disulfide bond are shownin Fig. 3BI, and positions for the attachment of MBB expectedfor a Cys-185–Cys-187 disulfide bond are shown in Fig. 3BII.After derivatization with MBB, excess of the latter was removedfrom the derivatized proteins while they were bound to anti-rhodopsin 1D4-Sepharose. Subsequently, the proteins wereeluted with the epitope nonapeptide. Step three involved diges-tion of the isolated derivatized proteins with proteinase K. Theresulting MBB-carrying peptide fragments were purified by theirselective binding to avidin-Sepharose (step four). Finally, after

Abbreviations: WT, wild type; RP, retinitis pigmentosa; TM, transmembrane; ID, intradiscal;PSB, protonated Schiff base; NEM, N-ethylmaleimide; MBB, maleimido-butyryl-biocytin;MALDI-TOF, matrix-assisted laser desorption ionization–time of flight.

See commentary on page 4819.

*Present address: Department of Pharmacology and Toxicology, Dartmouth MedicalSchool, 7650 Remsen, Hanover, NH 03755-3835.

†To whom reprint requests should be addressed. E-mail: khorana@mit.edu.

‡Amino acid substitutions in rhodopsin mutants are designated by using single letter codefor the original amino acid, followed by sequence number, and the single letter code forthe substituted amino acid.

§This is paper 42 in the series “Stucture and Function in Rhodopsin.” Paper 41 is ref. 15.

The publication costs of this article were defrayed in part by page charge payment. Thisarticle must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.§1734 solely to indicate this fact.

4872–4876 u PNAS u April 24, 2001 u vol. 98 u no. 9 www.pnas.orgycgiydoiy10.1073ypnas.061632798

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

elution with excess biotin, the MBB peptides were subjected toMALDI-TOF mass spectrometry (step five).§

Materials and MethodsMaterials. The 11-cis retinal was a gift from R. Crouch (Uni-versity of South Carolina, Charleston, SC, and the NationalEye Institute, National Institutes of Health, Bethesda). Thedetergent dodecyl maltoside (DM) was from Anatrace(Maumee, OH). All other reagents were from Sigma. Anti-rhodopsin monoclonal antibody rhodopsin-1D4 was purifiedfrom myeloma cell lines provided by R. S. Molday (Universityof British Columbia, Vancouver) and was coupled to cyanogenbromide-activated Sepharose 4B (Amersham Pharmacia) asdescribed (20). The buffers used were buffer A [NaCl, 137mMyKCl, 2.7 mMyKH2PO4, 1.8 mMyNaH2PO4, 10 mM (pH7.2)], buffer B (buffer A plus 0.05% DM), buffer C [2 mMNaH2PO4/Na2HPO4 (pH 6.0)y0.05% DM], and buffer D(buffer C plus 150 mM NaCl). All buffers were purged withargon.

Methods. Construction of the mutant opsin expression plasmids,transient expression in COS-1 cells, purification of the expressed

proteins, and separation of folded and misfolded opsins, havebeen described elsewhere in detail (14).

Derivatization of Sulfhydryl Groups in Free Cysteines in DenaturedRhodopsin with NEM. Separated, correctly folded rhodopsin mu-tants and the misfolded opsins (about 10 mg) in 300 ml of elutionbuffer C (containing epitope nonapeptide) or buffer D (con-taining epitope nonapeptide) were treated with NEM (100 mM)in the presence of 0.5% SDS for 3 h at room temperature. Theextent of derivatization with NEM was determined by titrationwith 49-49dithiodipyridine (21, 22). The reaction mixtures werethen concentrated to 100 ml by using Centricon 30 (Millipore),and the elution nonapeptide was removed by gel filtration byusing Sephadex G50 medium (Pharmacia). The NEM-derivatized proteins were then bound to 1D4-Sepharose suspen-sion by nutating for 2 h at 4°C.

Reduction of the Disulfide Bond in NEM-Derivatized Opsins or MutantRhodopsins with DTT and Derivatization of the Resulting SulfhydrylGroups with MBB. The NEM-derivatized protein [while bound to therhodopsin-1D4-Sepharose column (step 1)] was washed with 10 mlof buffer C. Buffer C (1 ml), containing 0.1% SDS and 10 mMDTT, was added to the suspension of the Sepharose beads, and themixture was incubated at 20°C for 60 min. The beads in a columnwere then washed with 5 ml of buffer C. Buffer B, containing 0.1%SDS and 0.1 mM MBB, was added to the suspension and nutatedend-over-end for 2 h at 20°C. After washing the beads with bufferB (20 ml), the MBB-derivatized protein was eluted from theSepharose beads with the epitope nonapeptide.

Digestion with Proteinase K. The MBB-derivatized protein wasdigested at room temperature for 6 h (1 h for the mutants H211Pand L125R) in 500 ml buffer B by using 1 mg Proteinase K

Fig. 1. A secondary structure model of rhodopsin showing the cytoplasmic, TM, and the ID domains. The 10 cysteines are shown in yellow with black circles;the three ID cysteines are highlighted by bolder circles. Amino acid sequences adjoining the three cysteines of interest (for identification of the disulfide bonds)are color coded for each cysteine: red for Cys-110, green for Cys-185, and blue for Cys-187. Native rhodopsin contains a disulfide bond between Cys-110 andCys-187 (indicated by dashed line). The RP mutants studied here are located in the TM domain (shown in blue boxes): G89D (helix II), L125R (helix III), A164V (helixIV), and H211P (helix V).

Fig. 2. Structure of maleimido-butyryl-biocytin (MBB).

Hwa et al. PNAS u April 24, 2001 u vol. 98 u no. 9 u 4873

BIO

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

(Sigma). The reaction was terminated by the addition of 0.1 mMPMSF.

Selective Binding of MBB-Containing Peptides to Avidin–Agarose andSubsequent Elution with Biotin. The proteinase K digestion mixturewas added to an avidin–agarose column equilibrated with bufferA and nutated end-over-end for 2 h at 20°C. The suspension waswashed with water, and the MBB-derivatized peptides wereeluted with a biotin solution (1 mgyml) in water. The water wasevaporated by using a speed vac, and the residue of MBB-derivatized peptides was dissolved in 5 ml of acetonitrileywater(1y1) containing 0.1% TFA.

MALDI-TOF of MBB-Labeled Peptides. One microliter of the sample(1–10 pM) was analyzed by mass spectrometry with a Voyagerinstrument (PerSeptive Biosystems, Framingham, MA). Thesample was transferred to the sample plate and allowed to dryby evaporation. The sample was then overlayed with 0.5 ml of thematrix compound (a-cyano-4-hydroxy cinnamic acid) at a con-centration of 10 mgyml. The program, MS PRODUCT(http:yyprospector.ucsf.eduyucsfhtml3.4ymsprod.htm), wasused to calculate the predicted monoisotopic and average myzto identify the peptides containing MBB-labeled cysteines. Ex-periments for WT and all of the mutant rhodopsins wererepeated at least three times.

ResultsWT Rhodopsin Contains a Cys-110–Cys-187 Disulfide Bond. The strat-egy (Fig. 2) was first applied to WT rhodopsin. As seen in Fig.4AI, the major experimentally observed myz of 1,575 corre-sponded to the peptide VFGPTGCNLE 1 MBB (expectedaverage, myz 1,574). This result showed Cys-110 as one of thecysteines participating in the disulfide bond. The second majorion, detected at myz 1,271, corresponded to CGIDYY 1 MBB(expected, myz 1,271). This result confirmed Cys-187 as thesecond cysteine involved in the disulfide bond with Cys-110.

Correctly Folded Portions of Mutants G89D (TM Helix II) and A164V (TMHelix IV) Contain a Cys-110–Cys-187 Disulfide Bond. Next, the fourRP mutants in different TM helices (Fig. 1) were studied.Previously, the mutants L125R (TM helix III) and H211P (TMhelix V) were found to cause complete misfolding, whereas

G89D (TM helix II) and A164V (TM helix IV) caused partialmisfolding (14). In the latter two cases, correctly folded andmisfolded portions of the expressed proteins were separated asdescribed elsewhere (10, 14). Mass spectrometric data obtainedwith the correctly folded portion of the mutant G89D are shownin Fig. 4AII. The two main ions seen, with experimental myz1,280 and myz 853, corresponded to the peptides, CNLEGF 1MBB (expected myz 1,280 for one Na1 and one K1 adduct) andCGI 1 MBB (expected myz 852 for one Na1 adduct). Thefinding of these peptides identified the presence of a Cys-110–Cys-187 disulfide bond in the correctly folded portion of mutantG89D. In the correctly folded portion of the mutant A164V (Fig.4AIII), the peptide (myz 1,085; GPTGCN 1 MBB) identifiedfrom the experimentally observed ion (expected myz 1,086)showed Cys-110 involvement in the disulfide bond. The ions atmyz 1,457 and myz 1,489, identified, respectively, SCGIDYYT1 MBB (expected myz 1,459) and CGIDYYTP 1 MBB (ex-pected myz 1,491 for Na1 adduct) demonstrating as the secondcysteine. (The ion at myz 777 could arise from labeling of eitherCys-187 or Cys-110, corresponding to peptide CG 1 MBB or GC1 MBB, respectively.) Thus, the correctly folded mutant,A164V, like the WT protein, contains a Cys-110-Cys-187 disul-fide bond.

Misfolded Rhodopsin RP Mutants Contain a Cys-185–Cys-187 DisulfideBond. Misfolded mutants H211P and L125R. Mass spectrometricstudies of the completely misfolded proteins from H211P andL125R are shown in Fig. 4B, I and II, respectively. In thespectrum of H211P, one ion was observed at myz 1,061, and wasassigned to the Cys-185 containing peptide GMQCS 1 MBB(expected myz 1,063). Two other signals observed in the samemutant, at myz 1,482 and myz 1,774, were identified as theCys-187-containing peptides, SCGIDYYT and CGIDYYTPHE(shown in blue in Fig. 4B). Further, the signals observed, myz1,711 and myz 1,998, corresponded to peptides CSCGID andCSCGIDYY, both Cys-185 and Cys-187 derivatized with MBB.In the spectrum of L125R, the ions seen with experimental myz862 and myz 945 corresponded to the Cys-187-containing pep-tides, SCG (expected myz 864 for the Na1 and K1 adduct) andCGID (expected myz 945), respectively. The ions at myz 1,240and myz 1,314 were assigned to the Cys-185-containing peptides,PEGMQC 1 MBB (expected myz 1,240 with K1 adduct) and

Fig. 3. Derivatizations as expected for ID cysteines in correctly folded (A) and misfolded (B) rhodopsins. The disulfide linkage in correctly folded A rhodopsinis between Cys-110 and Cys-187, as identified by mass spectrometry (Fig. 4). Thus, steps 1 and 2 of the strategy in the identification of disulfide bonds (see above)produced rhodopsin NEM-derivatized at Cys-185, whereas Cys-110 and Cys-187 were derivatized with MBB. In misfolded rhodopsin B, the disulfide bond couldeither be between Cys-110 and Cys-185 (I) or between Cys-185 and Cys-187 (II). Experimentally, the disulfide bond was found between Cys-185 and Cys-187 (Fig.4). Thus, Cys-110 was NEM-derivatized, whereas Cys-185 and Cys-187 were MBB-derivatized (II). The color code for the amino acid residues adjacent to each ofthe cysteines is the same as in Fig. 1. The derivatized products were then subjected to steps 3–5 (above), yielding the MBB-derivatized peptides that were analyzedby MALDI-TOF (Fig. 4).

4874 u www.pnas.orgycgiydoiy10.1073ypnas.061632798 Hwa et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

IPEGMQC 1 MBB (expected myz 1,315), respectively. Thus,the completely misfolded mutants H211P and L125R bothcontain a Cys-185–Cys-187 disulfide bond.

Misfolded portions of the proteins from G89D and A164V. Massspectrometric data on the misfolded portions of G89D and

A164V are shown in III and IV, respectively, of Fig. 4B. The ionsobserved at myz 1,124 and at myz 1,165 in the spectrum of themisfolded portion of G89D were both assigned to the Cys-185-containing peptide EGMQC 1 MBB with expected myz 1,127for the Na1 adduct and myz 1,165 for the Na1yK1 adduct. The

Fig. 4. MALDI-TOF Analysis of peptide fragments generated from the derivatized correctly folded (A) and misfolded (B) rhodopsins. The adjoining peptidesequences expected for the disulfide bonds are highlighted at the top in A and B. Rhodopsins were derivatized at cysteines as in the strategy outlined above.Because step 4 of the strategy selects only MBB-labeled peptides, all peptide assignments shown here include the mass for the MBB adducts of cysteines. They axis shows intensity in arbitrary units. The x axis shows myz in the range of 750–2,000. (A) Mass spectrometric results for WT rhodopsin and correctly foldedportions of the A164V and G89D mutants. The observed signals in the mass spectra correspond to MBB-labeled Cys-110 peptides (red) and Cys-187 peptides (blue).(B) Mass spectrometric results for the misfolded portion of A164V and G89D (III and IV) and of completely misfolded H211P and L125R (I and II). Only peptidescorresponding to a Cys-185–Cys-187 disulfide bond were observed. MBB-labeled Cys-185 containing peptides are shown in green, and MBB-labeled Cys-187peptides are shown in blue.

Hwa et al. PNAS u April 24, 2001 u vol. 98 u no. 9 u 4875

BIO

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

signals seen at myz 1,092 and myz 1,379 corresponded to theCys-187-containing peptides SCGID 1 MBB and SCGIDYY 1MBB. The signal at myz 805 could be caused by either theCys-185-containing peptide, CS 1 MBB, or the Cys-187-containing peptide, SC 1 MBB.

In the spectrum of the misfolded portion of mutant A164V,two signals were seen at myz 942 and myz 982. These corre-sponded to the peptides MQC 1 MBB (expected myz 941 for aNa1 adduct) and CGID (expected myz 983 for a K1 adduct).These two peptides identified a Cys-185–Cys-187 disulfide bondin the misfolded portion of mutant A164V. Thus, the misfoldedportions of the proteins from both G89D and A164V contain aCys-185–Cys-187 disulfide bond.

DiscussionMethods for the identification of disulfide bonds in water-solubleproteins were established in the 1960s (for example, ref. 23).However, integral membrane proteins have so far proven in-tractable for corresponding chemical-structural studies. Indeed,it is for this reason that the wealth of information now availableon primary structures of receptors, membrane transporters, andion channels has been made possible only by the development oftechnology for the sequencing of the corresponding genes. In thepresent work, we have developed a strategy for direct identifi-cation of disulfide bonds in correctly folded rhodopsin and itsmutants, and misfolded nonretinal binding opsins. The methodshould be applicable generally to corresponding studies ofintegral membrane proteins.

The strategy was first applied to identification of the nativedisulfide bond in rhodopsin. The conclusion that rhodopsincontains a Cys-110–Cys-187 disulfide bond had been arrived at,albeit indirectly, in a number of earlier studies (16–18). Directidentification of this disulfide bond as now accomplished is ofcentral significance in the superfamily of G protein-coupledreceptors (GPCRs). A disulfide bond at positions equivalent toCys-110 and Cys-187 in rhodopsin is now known to be conserved

in most of the GPCRs (ref. 24; see also, www.gcrdb.uthscsa.edu).This conservation implies the importance of this disulfide bondin a unitary mechanism for the activation of GPCRs (25). Otherstudies demonstrated misfolding in rhodopsin mutants preparedby designed mutagenesis (6–8), as well as in naturally occurringRP mutants (9, 10). Methods were developed for the separationand characterization of the misfolded opsin, and the conclusionwas drawn that misfolding caused by RP mutations in the IDdomain was caused by the formation of an abnormal disulfidebond (8). More recently, studies described the unequivocalidentification of this disulfide bond as that between Cys-185 andCys-187 (11, 12). In the present work, the abnormal disulfidebond present in four misfolded opsins resulting from RP muta-tions in TM helices II–V (Fig. 1) has been identified in every caseas that between Cys-185 and Cys-187. The total results nowfirmly establish that the packing of the seven helices in the TMdomain and folding in the ID domain are coupled (14).

Finally, it is noted that the Cys-185–Cys-187 disulfide bondbetween cysteines separated only by the single Ser-186 is unusualand, presumably, results from strain introduced in the ID tertiarystructure from the defective packing of the TM helices in RPmutants. There is a precedent for the presence of a disulfidebond with identical sequence, Cys-Ser-Cys, in mengovirus coatprotein (26, 27).

We are deeply indebted to Dr. Kevin Ridge (Center for AdvancedResearch in Biotechnology, Rockville, MD), who originally suggestedthe basic outline for the strategy developed in this article. We are gratefulto Professor U. L. RajBhandary (Massachusetts Institute of Technolo-gy), Dr. Ivan Corriera (Whitehead Institute), and Professor Jane S.Richardson (Duke University) for valuable discussions. We acknowledgethe enthusiastic assistance of Ms. Judy Carlin in the preparation of themanuscript. This work was supported by National Institutes of HealthGrant GM28289 and National Eye Institute Grant EY11717 (to H.G.K.).J.K.-S. was the recipient of a Howard Hughes Predoctoral Fellowship,and J.H. was the recipient of a Howard Hughes Medical InstitutePhysician Postdoctoral Fellowship.

1. Berson, E. L. (1993) Invest. Opthalmol. Visual Sci. 34, 1659–1676.2. Dryja, T. P. & Berson, E. L. (1995) Invest. Opthalmol. Visual Sci. 36, 1197–1200.3. Sung, C. H., Davenport, C. M., Hennessey, J. C., Maumenee, I. H., Jacobson,

S. G., Heckenlively, J. R., Nowakowski, R., Fishman, G., Gouras, P. & Nathans,J. (1991) Proc. Natl. Acad. Sci. USA 88, 6481–6485.

4. Macke, J. P., Davenport, C. M., Jacobson, S. G., Hennessey, J. C., Gonzalez-Fernandez, F., Conway, B. P., Heckenlively, J., Palmer, R., Maumenee, I. H.,Sieving, P., et al. (1993) Am. J. Hum. Genet. 53, 80–89.

5. Inglehearn, C. F., Keen, T. J., Bashir, R., Jay, M., Fitzke, F., Bird, A. C.,Crombie, A. & Bhattacharya, S. S. (1992) Hum. Mol. Genet. 1, 41–45.

6. Doi, T., Molday, R. S. & Khorana, H. G. (1990) Proc. Natl. Acad. Sci. USA 87,4991–4995.

7. Anukanth, A. & Khorana, H. G. (1994) J. Biol. Chem. 269, 19738–19744.8. Ridge, K. D., Lu, Z., Liu, X. & Khorana, H. G. (1995) Biochemistry 34,

3261–3267.9. Kaushal, S. & Khorana, H. G. (1994) Biochemistry 33, 6121–6128.

10. Liu, X., Garriga, P. & Khorana, H. G. (1996) Proc. Natl. Acad. Sci. USA 93,4554–4559.

11. Hwa, J., Reeves, P. J., Klein-Seetharaman, J., Davidson, F. & Khorana, H. G.(1999) Proc. Natl. Acad. Sci. USA 96, 1932–1935.

12. Kono, M., Yu, H. & Oprian, D. D. (1998) Biochemistry 37, 1302–1305.13. Garriga, P., Liu, X. & Khorana, H. G. (1996) Proc. Natl. Acad. Sci. USA 93,

4560–4564.

14. Hwa, J., Garriga, P., Liu, X. & Khorana, H. G. (1997) Proc. Natl. Acad. Sci. USA94, 10571–10576.

15. Cha, K., Reeves, P. J. & Khorana, H. G. (2000) Proc. Natl. Acad. Sci. USA 97,3016–3021.

16. Al-Saleeh, S., Gore, M. & Akhtar, M. (1987) Biochem. J. 246, 131–137.17. Karnik, S. S., Sakmar, T. P., Chen, H. B. & Khorana, H. G. (1988) Proc. Natl.

Acad. Sci. USA 85, 8459–8463.18. Karnik, S. S. & Khorana, H. G. (1990) J. Biol. Chem. 265, 17520–17524.19. Moffatt, J. G. & Khorana, H. G. (1961) J. Am. Chem. Soc. 83, 663–675.20. Oprian, D. D., Molday, R. S., Kaufman, R. J. & Khorana, H. G. (1988) Proc.

Natl. Acad. Sci. USA 84, 8874–8878.21. Klein-Seetharaman, J., Hwa, J., Cai, K., Altenbach, C., Hubbell, W. L. &

Khorana, H. G. (1999) Biochemistry 38, 7938–7944.22. Grassetti, D. R. & Murray, J. F., Jr. (1967) Arch. Biochem. Biophys. 119,

41–49.23. Brown, J. R. & Hartley, B. S. (1966) Biochem. J. 101, 214–228.24. Baldwin, J. M., Schertler, G. F. & Unger, V. M. (1997) J. Mol. Biol. 272,

144–164.25. Khorana, H. G. (2000) J. Biomol. Struct. Dyn. 11, 1–16.26. Krishnaswamy, S. & Rossmann, M. G. (1990) J. Mol. Biol. 211, 803–844.27. Luo, M., Vriend, G., Kamer, G. & Rossmann, M. G. (1989) Acta Crystallogr.

B 45, 85–92.

4876 u www.pnas.orgycgiydoiy10.1073ypnas.061632798 Hwa et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0